PSI - Issue 18

Ibrahim Al Zamzami et al. / Procedia Structural Integrity 18 (2019) 255–261 Author name / Structural Integrity Procedia 00 (2019) 000–000

258

4

For the butt and cruciform welded joints, two sets of thickness were used including 1 mm to run a fatigue test with a load ratio equal 0.1 and a thickness equal to 2 mm for fully reverse loading (R = -1). (The 2 mm thickness was used to perform a fully reversed loading to provide extra stiffness and prevent any bending effect under the compression loading.) As far as the lap joints are concerned, in order to keep a clear distance between the hydraulic grips and the weld seam, the resulting specimens are rather long and during the fully reversed loading, the bending effect was seen to be excessive and it was impossible to run the fatigue tests under R = -1. For this reason, a load ratio equal to 0.5 is used to investigate different loading levels. To prevent the bending of the main plate during the welding process, all the tee-welded joints were manufactured with a thickness equal to 2 mm and tested under load ratios equal to 0.1 and -1. 3. Nominal stress approach to estimate the fatigue life time of aluminium-to-steel welded joints The nominal stress method is by far the simplest and most commonly used approach for estimating the fatigue life of welded components. In this approach, the required design stresses are calculated according to classic continuum mechanics with reference to the nominal cross-sectional area. According to the definition of the nominal stress approach, the effect of the stress concentration arising from the weld tip should not be taken into account when calculating the nominal stresses. Because this effect is already considered in the reference fatigue design curves provided by the Standards and Codes of Practice (such as the International Institution of Welding (the IIW) (Hobbacher 2007), EC9 (Eurocode9 2011) and EC3 (Eurocode3 2005)). Therefore, selecting an appropriate design curve is essential to ensure an accurate fatigue design is achieved (Niemi 1995, Susmel 2009, Djavit & Strande 2013, Susmel 2014). To date, there is no recommendation guide for assessing the fatigue of aluminium-to-steel welded joints. However, in a recent study (Al Zamzami 2018), it was observed that the static fracture of aluminium-to-steel welded joints always occurred in the heat-affected zone of the aluminium plate. In addition, the fatigue experimental investigation revealed that the crack always initiates and propagates at the interface between the weld tip and the aluminium plate as shown in Figure 2. This shows that the onset of cracking was favoured by the localized stress concentration phenomenon that occurred in the weld bead region.

(a)

(b)

(c)

Fig. 2. Fatigue failure of (a) butt-welded (b) lap-welded (c) cruciform welded joints.

Clearly, the static and fatigue experimental investigations show that aluminium-to-steel welded joints behave very similarly to conventional welded joints. Therefore, this type of connection can be designed against fatigue loading by directly using the nominal stress approach along with the corresponding design curves recommended by the IIW, EC9 and EC3.